COMP408 - Parallel and Distributed Computing

246H.Gao et al.view,and then it incorporates different views by seeking optimal combination of multiple kernels of different views.(2)Subspace learning based approach.It obtains a latent consensus subspace shared by multiple views and cluster the instances on the latent subspace.There are many research works in this cat-egory,including CCA-based methods[3],spectral graph based methods[4–6], matrix factorization based methods[7,8].(3)Ensemble learning based approach.[9]takes a decision in each individual view separately and then combines all deci-sions of distinct views to establish a consensus decision by determining cluster agreements/disagreements.Traditional research assumes data are complete in all views.However,in many real applications,parts of instances are not available in some views.For example,in a news story clustering task,articles are collected from different on-line news sources.Only a part of news are reported in all views.No single source includes all news.Another example is image clustering.Images are based on multiple visual and textual features.Some images have only a fraction of visual or textual feature sets.Recently,a few attempts have been made for multi-view clustering with incomplete views.Thefirst work to deal with incomplete view clustering was pro-posed in[10].It uses one view’s kernel representation as the similarity matrix and complete the incomplete view’s kernel using Laplacian regularization.However, this approach requires that there exists at least one complete view containing all the instances.Shao et al.[11]relax the above constraint.They collectively complete the kernel matrices of incomplete datasets by optimizing the align-ment of shared instances of the datasets.Furthermore,a clustering algorithm is proposed based on the kernel canonical correlation analysis.However,this app-roach focus on two view problem.It can not exploit relation among more than two views.Li et al.proposed a Partial view clustering algorithm(PVC)[12]. Based on non-negative matrix factorization(NMF),PVC works by establishing a latent subspace where the instances corresponding to the same example in different views are close to each other.PVC concentrates on two views problem. Extending PVC to more views suffers from computational problem.In most recently,Shao et al.developed an incomplete view clustering algorithm(MIC) [13].MIC handles the situation of more than two incomplete views.With joint weighted non-negative matrix factorization,it learns a L2,1regularized latent subspace for multiple views.With mean value imputation initialization,MIC gives lower weights to the incomplete instances than the complete instances. During optimization,MIC pushes multiple views towards a consensus matrix iteratively.But,there are some limitations about MIC.It converges slowly and contains too many parameters,which makes it difficult to operate.Moreover, both PVC and MIC are NMF based method.Both of them inherit the limita-tions of NMF:(1)It cannot well deal with data with negative feature values. while in many real applications,the non-negative constraints can not be satisfied.(2)It is essentially linear,and thus cannot disclose non-linear structures hidden in data,which limits its learning ability.(3)It only deals with feature values, while in some applications we know the similarities(relationships)of instancesIncomplete Multi-view ClusteringHang Gao(B),Yuxing Peng,and Songlei JianScience and Technology on Parallel and Distributed Processing Laboratory, National University of Defense Technology,Changsha410073,China{gao.hang,pengyuxing,jiansonglei}@Abstract.Real data often consists of multiple views(or representa-tions).By exploiting complementary and consensus grouping informationof multiple views,multi-view clustering becomes a successful practice forboosting clustering accuracy in the past decades.Recently,researchershave begun paying attention to the problem of incomplete view.Gener-ally,they assume at least there is one complete view or only focus on twoview problems.However,above assumption is often broken in real tasks.In this work,we propose an IVC algorithm for clustering with more thantwo incomplete pared with existing works,our proposed algo-rithm(1)does not require any view to be complete,(2)does not limitthe number of incomplete views,and(3)can handle similarity data aswell as feature data.The proposed algorithm is based on the spectralgraph theory and the kernel alignment principle.By aligning projec-tions of individual views with the projection integration of all views,IVC exchanges the complementary grouping information of incompleteviews.Consequently,projections of individual views are made completeand thereby resulting the consensus with accurate grouping information.Experiments on synthetic and real datasets demonstrate the effectivenessof IVC.Keywords:Multi-view clustering·Incomplete view clustering·Spectral clustering1IntroductionMany datasets in real world are naturally comprised of heterogeneous views(or representations).Clustering with such type of data is commonly referred to as multi-view Clustering.With the assumption of complementary data representa-tion and consensus decision of clusterings,multi-view clustering has the potential to dramatically increase the learning accuracy over single view clustering[1].The main problem in multi-view clustering is how to integrate grouping information of individual views.Existing works can be roughly classified into three categories.(1)Multi-kernel learning based approach.The most representative work of this category is Multi-kernel Kmeans[2].Itfirst uses kernel representation for each c IFIP International Federation for Information Processing2016Published by Springer International Publishing AG2016.All Rights ReservedZ.Shi et al.(Eds.):IIP2016,IFIP AICT486,pp.245–255,2016.DOI:10.1007/978-3-319-48390-025。

CONCURRENCY AND COMPUTATION:PRACTICE AND EXPERIENCEConcurrency Computat.:Pract.Exper.2005;17:117–141Published online in Wiley InterScience().DOI:10.1002/cpe.931Performance technology forparallel and distributedcomponent softwareA.Malony1,∗,†,S.Shende1,N.Trebon1,J.Ray2,R.Armstrong2,C.Rasmussen3and M.Sottile31Department of Computer and Information Science,University of Oregon,Eugene,OR97403,U.S.A.2Sandia National Laboratory,Livermore,CA94551,U.S.A.3Los Alamos National Laboratory,Advanced Computing Laboratory,Los Alamos,NM87545,U.S.A.SUMMARYThis work targets the emerging use of software component technology for high-performance scientific parallel and distributed computing.While component software engineering will benefit the construction of complex science applications,its use presents several challenges to performance measurement,analysis,and optimization.The performance of a component application depends on the interaction(possibly nonlinear) of the composed component set.Furthermore,a component is a‘binary unit of composition’and the only information users have is the interface the component provides to the outside world.A performance engineering methodology and development approach is presented to address evaluation and optimization issues in high-performance component environments.We describe a prototype implementation of a performance measurement infrastructure for the Common Component Architecture(CCA)system.A case study demonstrating the use of this technology for integrated measurement,monitoring,and optimization in CCA component-based applications is given.Copyright c 2005John Wiley&Sons,Ltd.KEY WORDS:component software;performance;parallel;distributed;optimization1.INTRODUCTIONThe power of abstraction has played a key role throughout the history of scientific computing in managing the growing complexity of scientific problem solving.While the evolution of software abstractions and the technology that supports them has helped to address the challenges of scientific∗Correspondence to:A.Malony,Department of Computer and Information Science,University of Oregon,Eugene,OR97403, U.S.A.†E-mail:malony@Contract/grant sponsor:U.S.Department of Energy;contract/grant numbers:DF-F603-01ER25501and DE-F602-03ER25561Copyright c 2005John Wiley&Sons,Ltd.Received2December2003 Revised16February2004 Accepted25March2004118 A.MALONY ET AL.application development,it has at times been in conflict with the ability to achieve high performance. On the one hand,abstraction in software systems further distance the scientific application developer from the range of sources of performance behavior and possible performance problems.On the other hand,support for performance observation and analysis has been poorly integrated in software systems, making performance evaluation a more difficult task.As both the power and the complexity of scientific software environments and computing systems mutually advance,it is imperative that technology for performance evaluation and engineering keep pace and become part of an integrated software and systems solution.The software challenges of building large-scale,complex scientific applications are beginning to be addressed by the use of component software technologies(see,e.g.,[1–3]).The software engineering of scientific applications from components that can be‘plugged’together will greatly facilitate construction of coupled simulations and improve their cross-platform portability.However, the success of scientific component software will depend in the end on the ability to deliver high-performance solutions.Scientific components are more complex and diverse than typical software components or libraries,in their scale,execution modes,programming styles and languages, and system targets.Performance technology that lacks robustness,portability,andflexibility will inevitably prove incapable of addressing the software and platform integration requirements required for performance observation and analysis.Intra-component performance engineering addresses problems of individual component performance characterization,analysis,modeling,and adaptation. Inter-component performance engineering provides local and overall awareness of application performance and facilities to access that performance information for the application to utilize. Most importantly,performance engineering technology should be compatible with the component engineering methodologies and frameworks used to develop applications,or it will be neither routinely nor effectively applied by component and application developers.Our research work on performance technology for component software is defined by three objectives.Thefirst is a methodological and operation model for intra-and inter-component performance engineering.Here,we define how the performance engineering technology will integrate with the component architectures and frameworks being considered for scientific computing. The Common Component Architecture(CCA)Forum[4]is specifying component software extensions and infrastructure to address problems of parallel component interfaces,scientific data exchange, and cross-language interoperability.We chose the CCA specification as our reference archetype. The second objective is the development of technology to implement the methods and techniques required for intra-and inter-component performance engineering in the context of existing scientific component efforts.Our target audience for the performance engineering technology we are creating are the framework and application developers using the CCA specification.Specifically,we are integrating our T AU performance system[5]with the CCA software,such as the Scientific Interface Definition Language(SIDL)[6],the Babel component interface toolkit[7],and the CCAFFEINE framework[1]. Ourfinal objective is the application of the model and technology for performance engineering to real component-based scientific computing environments.The goal here is to demonstrate both the capability and utility of our performance engineering ideas.The three objectives above are covered in this paper.Section2discusses the use of component technology for scientific computing and motivates the general requirements for performance engineering.The functional operation of the CCA specification is also described.Our conceptual model for performance engineering of component software is presented in Section3.Here we describe Copyright c 2005John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2005;17:117–141PERFORMANCE TECHNOLOGY FOR COMPONENT SOFTWARE119 a high-level methodology to technology development for intra-and inter-component performance engineering.Section4then considers the implementation of this technology in the context of the CCA software environment.Section5presents an application of our work with an emphasis on CCA performance modeling and optimization.Here we demonstrate the use of a CCA performance interface in measurement experiments to construct empirical performance models.We also show how, when the application runs,an optimizing component utilizes the performance application programmer interface(API)to gather statistics about the running application and decides which of the sets of similar components to choose for optimal performance.Our work has been targeted primarily towards high-performance computing(HPC)environments. However,it naturally extends to Grid computing both in terms of application of component technologies and the requirements for performance engineering.We briefly discuss the extensions and the issues that arise as part of the conclusions Section6.PONENT TECHNOLOGY FOR SCIENTIFIC COMPUTINGComponent technology extends the benefits of scripting systems and object-oriented design to support reuse and interoperability of component software,transparent of language and location[8].A component is a software object that implements certain functionality and has a well-defined interface that conforms to a component architecture defining rules for how components link and work together. The term component framework is used to designate a specific component architecture implementation. Component technology offers many advantages to scientific computing since it allows domain-level knowledge to be encapsulated in component building blocks that can easily,hopefully efficiently,be used in application development,removed from the concerns of how the component was developed or where it resides.As a result,scientists can focus their attention to overall application design and integration.Unfortunately,the three most widely-used component standards(CORBA[9],COM/DCOM[10], Java Beans[11])are ill-suited to handle high-performance scientific computing due to a lack of support for efficient parallel communication,insufficient scientific data abstractions(plex numbers), and/or limited language interoperability[12].Furthermore,often the software does not run on the systems scientists use,or simply runs too slow for their applications.ATo overcome some of the limitations of standard component software approaches for scientific computing,the CCA Forum[4]was started in1997to define the standard foundations of scientific component architecture and to facilitate the adoption of CCA tools and technologies.In addition,the U.S.Department of Energy(DOE)established the Center for Component Technology for Terascale Simulation Software(CCTTSS)[13]for purposes of developing the CCA software infrastructure and demonstrating its use for complex scientific simulations.Component programming,much like object-oriented programming,provides a model for constructing software such that units of code(components and objects)expose a‘public’interface to the outside while hiding their internal implementation ponents extend the object model by allowing components to dynamically discover and expose interface information,something that is Copyright c 2005John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2005;17:117–141120 A.MALONY ETAL.Figure1.Two CCA components:One uses a‘P’port provided by the other.statically determined at compilation time in most object-oriented languages.Fundamentally,the CCA is a specification of the component programming pattern and the interface the components see to the underlying support substrate,or framework.The CCA allows components to describe their interfaces in the Scientific Interface Definition Languaged(SIDL)[6].Like the IDL used by CORBA,the interfaces are defined in a language independent manner and are not bound to the source code or compiled binary of a component.The IDL simply describes the public interface so that external parties can discover what services are available and how they must be called.In the CCA,a component is defined as a collection of ports,where each port represents a set of functions that are publicly available.A port is described using SIDL,and some form of wrapper exists in the implementation to map the SIDL interface to that of the implementation language.From the point of view of a component,there are two types of ports.Those that are implemented by a component are known as provides ports,and other components may connect to and use them.Other ports that a component will expect to be connected to and call are known as uses es and provides ports are connected together as shown in Figure1.The act of connecting components is referred to as component composition.When a component is instantiated and allowed to execute,it registers the provides and uses ports with the underlying framework.This information allows external components to discover what ports or interfaces are available,and ensures that expected relationships between components are fulfilled, before allowing execution.Port discovery is a service provided by the framework and is actually just another port that a component can connect to.For instance,a component can obtain a list from the framework of all components providing a specific interface or port.The component could then connect to each of the ports in the list in an iterative fashion and make calls on the methods within the connected port.Iterative access to a common set of interfaces presupposes that such a set exists. One of the primary benefits of component-based programming is the adoption of common interfaces for domain-specific purposes.This allows different teams of individuals to develop components based on this‘standard’component API.This,in turn,allows users of these components to pick and choose the particular component that bestfits their needs.The CCA architecture is the operational foundation for the implementation of CCA-compliant component frameworks,but it does not by itself directly specify how requirements of high-performance component systems are met.Rather,the intent of the CCA specification is to allow the preservation of performance.This takes four primary forms.First,the performance of the functional‘core’of a component(e.g.linear equation solver)should not suffer because it is implemented within a component.In general,this demands support for all core languages(e.g.C,C++,and Fortran77, 90and95),and support for all forms of parallel execution that the core software may employ. Copyright c 2005John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2005;17:117–141PERFORMANCE TECHNOLOGY FOR COMPONENT SOFTWARE121Second,component communication mechanisms should not be dictated but be selectable based on what best suits encapsulated functionality and component–component proximity(e.g.same memory space,same cluster,local network or wide-area).Third,parallelism between components should not be restricted.This regards primarily support for parallel communication between components, but also relates to parallel component computing paradigms,as in the Single Component Multiple Data(SCMD)model[1],MxN coupling[14],and the macro-level dataflow Uintah Computational Framework(UCF)[15].Lastly,it should be possible to select component instances and configure component compositions for performance purposes,both prior to and during execution.2.2.Performance engineering and component softwareThe above important performance-related aspects of component-based scientific computing motivate fundamental performance engineering questions that are examined specifically in our research.•How are the performance of components and component compositions evaluated?•How can this evaluation be done robustly given the diversity of component(component composition)types and implementations?•How are component performance data represented and made available to the component framework as a whole?•How can the performance of component compositions be modeled?•What restrictions does a component approach place on performance engineering,and can the component architecture,framework infrastructure,and technologies be leveraged to implement and deliver performance engineering support more effectively?•Will the integration of performance engineering technology lead to a quantifiable improvement in performance of component-based scientific computing and how will this be demonstrated?•Is it possible to develop a component performance engineering solution that is general purpose and can be applied in many scientific computing contexts?We address these questions from the perspective of the technology to deliver performance-engineered solutions for scientific computing environments.2.3.Related workSince commercial component models are targeted mostly at serial computing environments, the design of performance methods and metrics for these software systems are not likely to account for critical requirements important to high-performance scientific computing such as memory hierarchy performance,data locality,orfloating point operation.Likewise,the distributed frameworks/component models(e.g.DCOM and CORBA)use commodity networking to connect components together,and lack consideration for high-performance network communication required for HPC.In a distributed environment,metrics such as round-trip time and network latency are often considered useful,while quantities such as bisection bandwidth,message-passing latencies and synchronization cost,which form the basis of much of the research in scientific performance evaluation, are left unaddressed.Copyright c 2005John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2005;17:117–141122 A.MALONY ET AL.However,despite the different semantics,several research efforts in these standards offer viable strategies in measuring performance.A performance monitoring system for the Enterprise Java Beans standard is described in[16].For each component to be monitored,a proxy is created using the same interface as the component.The proxy intercepts all method invocations and notifies a monitor component before forwarding the invocation to the component.The monitor handles the notifications and selects the data to present,either to a user or to another component(e.g.a visualizer component). The goal of this monitoring system is to identify hot spots or components that do not scale well.The Wabash tool[17,18]is designed for pre-deployment testing and monitoring of distributed CORBA systems.Because of the distributed nature,Wabash groups components into regions based on the geographical location.An interceptor is created in the same address space of each server object(i.e.a component that provides services)and manages all incoming and outgoing requests to the server.A manager component is responsible for querying the interceptor for data retrieval and event management.In the work done by the Parallel Software Group at the Imperial College of Science in London[19,20],the research is focused on Grid-based component computing.However,the performance is also measured through the use of proxies.Their performance system is designed to automatically select the optimal implementation of the application based on performance models and available resources.With n components,each having C i implementations,there is a total ofn i=1C i implementations to choose from.The performance characteristics and a performance modelfor each component is constructed by the component developer and stored in the component repository. Their approach is to use the proxies to simulate an application in order to determine the call-path. This simulation skips the implementation of the components by using the proxies.Once the call-path is determined,a recursive composite performance model is created by examining the behavior of each method call in the call-path.In order to ensure that the composite model is implementation-independent,a variable is used in the model whenever there is a reference to an implementation. To evaluate the model,a specific implementation’s performance model replaces the variables and the composite model returns an estimated execution time or estimated cost(based on some hardware resources model).The implementation with the lowest execution time or lowest cost is then selected and a execution plan is created for the application.3.PERFORMANCE ENGINEERING METHODOLOGYBuilding efficient scientific applications as hierarchical compositions of cooperating components depends significantly on having a thorough knowledge of component performance and component interactions.To make effective decisions concerning component configuration,deployment,and coupling,it is important to develop a performance engineering methodology that complements the component and application development and execution processes.Ideally,the methodology would be supported by performance technology(measurement,analysis,and modeling)that extends the programming and execution environment to be performance observable and performance aware. However,the diversity of component functionality and the complexity of component implementation (e.g.different languages,hardware platforms,and parallelism modes)challenge performance technology to offer(more)robust solutions.Our objective here is to define a consistent performance engineering model for components and component ensembles,and to implement that model in an integrated fashion throughout the component development and execution framework.Copyright c 2005John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2005;17:117–141PERFORMANCE TECHNOLOGY FOR COMPONENT SOFTWARE123Figure2.Performance engineered component design.ponent performance engineeringWhile component architectures facilitate the development of complex applications by allowing the independent creation of generic,reusable components and by easing their composition,application-level performance engineering requires knowledge about a component’s performance in different contexts and observation of the component’s performance behavior during execution.We view a performance-engineered component as having four constituent parts:•performance characterization and modeling(performance knowledge);•integrated performance measurement and analysis(performance observation);•runtime access to performance information(performance query);•mechanisms to alter component performance behavior(performance control).In Figure2,we show how a component’s design may be extended to support these performance engineering features.The intent is to keep the extended(performance engineered)design relatively consistent with the CCA model.We represent a component’s generic architecture by the middle(solid)box in thefigure. As shown,we distinguish between a component’s‘core’functionality and its variant sub-parts(both functional variants and code variants)that may be selected prior to or during component execution‡.‡This view is consistent with how components are regarded in several systems.(Note,a component’s ports can also have variants, but this is not drawn to reduce diagram complexity.)Copyright c 2005John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2005;17:117–141124 A.MALONY ET AL.A component’s generic design can be‘performance engineered’with support for performance knowledge and/or performance observation.Performance knowledge extensions(left(dashed)box) provide means to describe and store what is‘known’about the component’s performance.This can take the form of empirical characterizations in a performance database,as well as performance models captured from empirical studies or analytical performance analysis.Performance observation extensions(right(dashed)box)implement performance measurement and analysis capabilities that are used during component execution.In addition,support for querying the performance information and for effecting control based on performance feedback complement the performance knowledge and observation parts.To justify these extensions,let usfirst consider the use of performance knowledge in the performance engineering of component frameworks.The ability to save information about a component’s performance allows that knowledge to be used for performance-guided component selection,deployment,and runtime adaptation.However,the representation of performance knowledge must be in common forms and there must be standard means to access selective performance information.In Figure2,we identify a‘component performance repository’where the performance knowledge could be kept and queried within the component framework,similar in concept(and, likely,function)to the CCA component repository.In fact,we could view the performance knowledge extension as a component in its own right,as suggested in thefigure by the separation of the left (dashed)box from original component.In this way,the performance knowledge component(PKC) could provide PKC ports that give component-level access to the performance information both to other components within the framework(black arrowhead)as well as back to the original component (shaded arrowhead)whose performance it represents.These‘feedback’port connections would allow for both static(instantiation-time)and dynamic(runtime)component control.The PKCs could also be instantiated and used,in particular,as active components to build runtime composite performance models for resource allocation and scheduling.Similarly,let us consider justification for performance observation support in our performance engineered component model.The ability to observe a component’s execution time performance is important for two reasons.By definition,any empirically derived performance knowledge requires some form of performance measurement and analysis.However,it does not mean that this support necessarily must be integrated in the component’s design.The second reason is to monitor the performance of a component during execution and use that information in dynamic performance decision making.Here,integrated performance observation support(measurement and analysis) is key.As depicted in Figure2(block-filled double arrow),this integration requires component instrumentation(both core and variant),runtime measurement and data collection,and online and/or offline performance analysis.In this respect,we can view performance observation in the performance engineered component model as a functional extension of the original component design.To allow the component framework to query performance behavior,this extension could include new component methods and ports(black arrowhead).It is also useful for the original component to be able to access its own performance observations to make internal performance decisions.An interesting design issue here concerns how this functionality should be supported.Our view is to further generalize the model to regard observation support as encapsulated in a performance observation component(POC)that is tightly coupled and co-resident with the original component.Special POC‘provides’ports allow the original component to use optimized interfaces(shaded arrowhead)to access‘internal’performance observations.Copyright c 2005John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2005;17:117–141PERFORMANCE TECHNOLOGY FOR COMPONENT SOFTWARE 125}hierarchical component composition performance awareness monitoring}Figure 3.Hierarchical performance awareness monitoring for component composition.ponent composition performance engineering Our discussion above focused on the performance engineering model at the component level.Now we turn attention to component composition.The performance of component-based scientific applications depends as much (perhaps more)on the interplay of component functions and the computational resources used for component execution and interoperation,as it does on individual component performance.Because component operations and interactions can be complex and dynamic,and resource availability may be subject to variation,the management of component compositions throughout their execution becomes a critical element of their successful deployment and use.In general,this management can include resource assignment,component placement and scheduling,data communications control,and file storage and allocation.However,the issue for us is not necessarily to design the specific tools that will be part of the performance management solution,but to identify key technological capabilities needed to support the performance engineering of component ensembles.In practice,this distinction is blurred,admittedly due to the different infrastructure used for scientific component computing,ranging from wide-area Grids of heterogeneous resources to tightly-coupled large-scale computing systems.Thus,we restrict our attention to only two model concepts:performance awareness and performance attention .Performance awareness of component compositions relates to an ensemble view of performance,how this information is obtained,and how it can be accessed.As with components,performance engineering looks at composition performance knowledge and position performance knowledge can come from empirical as well as analytical evaluation,can utilize information provided at the component level,and can be stored in repositories for future review.Performance awareness extends the notion of component observation to ensemble-level performance monitoring .The idea is to associate monitoring components with levels of hierarchical component grouping,building upon component-level observation support.This is presented in Figure 3.These monitoring componentsCopyright c 2005John Wiley &Sons,Ltd.Concurrency Computat.:Pract.Exper.2005;17:117–141。

序号会议名称1ACM SIGCOMM: ACM Conf on Communication Architectures, Protocols & Apps2IEEE INFOCOM: The Conference on Computer Communications3IEEE International conference on communications4IEEE Globecom: Global Telecommunications Conference5IEEE ITC: International Test Conference6IEEE The International Conference on Dependable Systems and Networks7ACM MobiCom: International Conference on Mobile Computing and Networking8ACM SIGMETRICS: Conference on Measurement and Modeling of Computer Systems9MOBIHOC: ACM International Symposium on Mobile Ad Hoc Networking and Computing10IEEE International Conference on Distributed Computing Systems11IMC: Internet Measurement Conference12ICCV: IEEE International Conference on Computer Vision13CVPR: IEEE Conf on Comp Vision and Pattern Recognition14ECCV: European Conference on Computer Vision 15DCC: Data Compression Conference16ICML: International Conference on Machine Learning17NIPS: Neural Information Processing Systems 18ACM MM: ACM Multimedia Conference19IEEE ICIP: International conference on Image Processing20IEEE ICME: International Conference on Multimedia and Expo21IEEE VR:IEEE Virtual Reality22ACM VRST:ACM Virtual Reality Software and Technology23CGI:Computer Graphics International24ACL: The Association for Computational Linguistics 25COLING: International Conference on Computational Linguistics 26IEEE ICASSP: International Conference on Acoustics, Speech and Signal Processing 27IJCNLP: International Joint Conference on Natural Language Processing 28IEEE/ACM Design Automation Conference 29IEEE VLSI Test Symposium 30IEEE/ACM Design, Automation and Test in Europe 31IEEE Asian Test Symposium 32Ubicomp: International Conference on Ubiquitous Computing 33PerCom: IEEE International Conference on Pervasive Computing and Communications 34EUC: The IFIP International Conference on Embedded And Ubiquitous Computing 35ICPS: IEEE International Conference on PervasiveServices36SenSys，ACM Conference on EmbeddedNEtworked Sensor Systems37SECON, IEEE Communication Society Conference on Sensor and Ad Hoc COmmunications andNetworks,38MASS, IEEE International Conference on Mobile Ad hoc and Sensor Systems39The International Conference for High Performance Computing and Communication 40CLUSTER 4, IEEE Int’l Conf. on ClusterComputing,41HPDC-, th IEEE Int’l Symp. on High-Performance Distributed Computing, Honolulu.42 NASA/IEEE Conference on Mass Storage Systems an43SuperComputing：The International Conference for High Performance Computing andCommunications44IEEE Int'l Conf. on Cluster Computing45[ICDCS] International Conference on Distributed Computing Systems46[HPDC] IEEE Int'l Symp. On High Performance Distributed Computing47International Conference for High Performance Computing and Communications (IEEE/ACMSupercomputing Conference)48ACM International Conference on Supercomputing49IEEE International Parallel & DistributedProcessing Symposium50IEEE International Conference on ParallelProcessing 51IEEE International Conference on High Performance Computing52ACM SIGMETRICS International Conference on Measurement and Modeling of Computer Systems.53IEEE Annual Workshops on WorkloadCharacterization.54International Symposium on ComputerArchitecture(ISCA)55International Symposium on High Performance Computer Architecture(HPCA)56International Symposium on Microarchitecture (MICRO)57 FAST: USENIX Conference on File and Storage Technologies,58NASA/IEEE Conference on Mass Storage Systems and Technologies (MSST4)，59 SNAPI’4: International Workshop on Storage Network Architecture and Parallel I/Os, Antibes Juan-les-spins, French,60 IEEE SC: SC-High Performance Computing, Networking and Storage Conference61IEEE International Workshop on Networking, Architecture, and Storages（IWNAS）62IEEE International Conference on Autonomic Computing（ICAC）63Proceedings of the International Conference on Measurements and Modeling of Computer Systems64International Symposium on High Performance Computer Architecture(HPCA)65 [HPDC] IEEE Int'l Symp. On High Performance Distributed Computing66IEEE Int'l Conf. on Cluster Computing 67USENIX Annual Technical Conference68IEEE/ACM Int'l Symp. on Cluster Computing & the Grid69International Symposium on Computer Architecture（ISCA ）70International Symposium on Microarchitecture （MICRO）71HPCC：The International Conference for High Performance Computing and Communications72 IEEE International Conference on High Performance Computing73 Annual ACM International Conference on Supercomputing（ICS）74Symposium on Operating System Design and Implementation（OSDI）75 ACM Symposium on Operating Systems Principles （SOSP）76Proceedings of the 11th International Conference on Architectural Support for Programming Languages and Operating Systems （ASPLOS）77 Workshop on Hot Topics in Operating Systems （HOTOS）78Proceedings of the International Conference on Parallel Processing（ICPP）79 Annual IEEE Conference on Local Computer Networks（LCN）80International Conference on Distributed Computing Systems（ICDCS）81International Conference on Parallel and Distributed Computing, Applications and Technologies（PDCAT）82IEEE International Parallel and Distributed Processing Symposium（IPDPS）83ASPLOS: Architectural Support for Programming Languages and Operating Systems84CASES: International Conference on Compilers, Architectures and Synthesis for Embedded Systems85CODES: International Conference on Hardware Software Codesign86DAC: Annual ACM IEEE Design Automation Conference87ICFP: International Conference on Functional Programming88ICS: International Conference on Supercomputing89ICSE: International Conference on Software Engineering90ISCA: International Conference on Computer Architecture91ISMM: International Symposium on Memory Management92ISSTA: International Symposium on Software Testing and Analysis93LCTES: Language, Compiler and Tool Support for Embedded Systems94MICRO: International Symposium on Microarchitecture95OOPSLA: Conference on Object Oriented Programming Systems Languages and Applications96PLDI: Conference on Programming Language Design and Implementation97PODC: Annual ACM Symposium on Principles of Distributed Computing98POPL: Annual Symposium on Principles of Programming Languages99PPoPP: Principles and Practice of Parallel Programming100SIGMETRICS: Joint International Conference on Measurement and Modeling of Computer Systems101SIGSOFT: Foundations of Software Engineering102ASE - IEEE International Conference on Automated Software Engineering103CGO - International Symposium on Code Generation and Optimization104CLUSTER - IEEE International Conference on Cluster Computing106EURO-PDP - Euromicro Conference on Parallel, Distributed and Network-Based Processing107HPCA - International Symposium on High-Performance Computer Architecture108HPCS - Annual International Symposium on High Performance Computing Systems and109ICDCS - International Conference on Distributed Computing Systems110ICPADS - International Conference on Parallel and Distributed Systems105DATE - Design, Automation, and Test in Europe111IISWC - IEEE International Symposium on Workload Characterization112IPDPS - International Parallel and Distributed Processing Symposium113ISPASS - IEEE International Symposium on Performance Analysis of Systems and Software114PACT - International Conference on Parallel Architectures and Compilation Techniques115RTSS - IEEE Real-Time Systems Symposium116RTAS - IEEE Real-Time and Embedded Technology and Applications Symposium117SC - IEEE/ACM SC Conference118LCPC - International Workshop on Languages and Compilers for Parallel Computing119CC: International Conference on Compiler Construction120HiPEAC - International Conference on High Performance Embedded Architectures & Compilers121ECOOP - European Conference on Object-Oriented Programming122ESOP - European Symposium on Programming123Euro-Par - European Conference on Parallel Computing124SAS - International Static Analysis Symposium 125CAV - Computer Aided Verification126FASE - Fundamental Approaches to Software Engineering127TACAS - Tools and Algorithms for the Construction and Analysis of Systems128VMCAI - Verification, Model Checking and Abstract Interpretation 129ACL: The Association for Computational Linguistics 130ACM SIGIR: The ACM Conference on Researchand Development in Information Retrieval 131ACM SIGKDD: The ACM Conference on Knowledge Discovery in Databases and Data Mining 132WWW: The ACM International World Wide Web Conference133ACM SIGMOD: ACM SIGMOD Conf on Management of Data 134CIKM: The ACM Conference on Information and Knowledge Management 135COLING: International Conference on Computational Linguistics 136ICML: International Conference on MachineLearning137IEEE ICDM: International Conference on DataMining 138IJCAI: International Joint Conference on Artificial Intelligence139VLDB: The ACM International Conference onVery Large Data Bases140SIGGRAPH: ACM SIGGRAPH Conference 141EUROGRAPHICS： The Annual Conference of the European Association for Computer Graphics 142AAAI: American Association for Artificial Intelligence 143ACM Conference　on　Computer and Communications Security 144ACM SIGCOMM: Special Interest Group on Data Communications 145ACM SIGIR: The ACM Conference on Researchand Development in Information Retrieval 146ACM SIGKDD: The ACM Conference on Knowledge Discovery in Databases and Data Mining 147ACM SIGMETRICS: Conference on Measurement and Modeling of Computer Systems148ACM SIGMOD: ACM SIGMOD Conf on Management of Data 149ASPLOS: Architectural Support for ProgrammingLanguages and Operating Systems150CCGrid : IEEE/ACM International Symposium on Cluster Computing and the Grid 151CIKM: The ACM Conference on Information and Knowledge Management 152CLUSTER - IEEE International Conference onCluster Computing153CPM: Combinatorial Pattern MatchingSymposium 154FAST4: Third USENIX Conference on File and Storage Technologies,155Grid : IEEE/ACM International Workshop on Grid Computing 156HPC: IEEE International Conference for High Performance Computing 157HPDC: International Symposium on High Performance Distributed Computing 158ICDCS: IEEE International Conference on Distributed Computing Systems 159ICML: International Conference on Machine Learning 160ICWS: IEEE International Conference on WebServices161IEEE CSB: Computer Society Bioinformatics 162IEEE ICDM: International Conference on DataMining 163IEEE ICNP: International Conference on Network Protocols164IEEE ICON: IEEE International Conference on Networks 165IEEE INFOCOM: conference on computer communications 166IEEE IPCCC: International Performance Computing and Communications Conference 167IEEE SPIRE: The IEEE International Symposium onString Processing and Information Retrieval 168IJCAI: International Joint Conference on AI 169IMC: ACM-SIGCOMM/USENIX Internet Measurement Conference170International Symposium on High Performance Computer Architecture(HPCA)171IPDPS: IEEE International Parallel and Distributed Processing Symposium172ISMB: International conference on Intelligent Systems for Molecular Biology173MobiCom: ACM/IEEE Conference on Mobile Computing and Networking174MobiSys: The International Conference on Mobile Systems, Applications, and Services175OSDI: USENIX Symposium on Operating Systems Design and Implementation176PAKDD: Pacific-Asia Conference on Knowledge Discovery and Data Mining177PDCAT: International Conference on Parallel and Distributed Computing, Applications and Technologies178PKDD: Conference on Principles and Practice of Knowledge Discovery in Databases179SCC: IEEE International Conference on Services Computing180SDM: SIAM International Conference on Data Mining181SOSE: IEEE International Workshop on Service-Oriented System Engineering182USENIX Sec: USENIX Security Symposium183USENIX/ACM Symposium on Networked Systems Design and Implementation (NSDI)184VLDB: The ACM International Conference on Very Large Data Bases185WWW: The ACM International World Wide Web Conference186RAID International Symposium on Recent Advances in Intrusion Detection187IJCAI: International Joint Conference on Artificial Intelligence188VLDB: The ACM International Conference on Very Large Data Bases189ICML: International Conference on Machine Learning190PRICAI: Pacific Rim International Conference on Artificial Intelligence191IFIP ICIIP: IFIP International Conference on Intelligent Information Processing192NIPS: Neural Information Processing Systems193ISCA: International Symposium on Computer Architecture194International Symposium on Microarchitecture195HPCA：International Symposium on High Performance Computer Architecture196APCSAC: Asia-Pacific Computer Systems Architecture Conference 197ISLPED: International Symposium on Low PowerElectronics and Design198OSDI: Operation System Design &Implementation 199ASPLOS: Architecture Support for Programming Languages and Operation200ICCD: IEEE International Conference onComputer Design201DAC: Design Automation Conference 202IEEE/ACM International Conference on ComputerAided Design(ICCAD)203ASP-DAC: Asia and South Pacific DesignAutomation Conference204ISSCC: IEEE International Solid-State CircuitsConference205CICC: Custom Integrated Circuits Conference 206ESSCIRC: European Conference on Solid-StateCircuits207Symposium on VLSI Circuits208IEEE International ASIC/SOC Conference209Symposium on VLSI Technology210ASSCC: Asian Conference on Solid-State Circuits211MWSCAS: Midwest Symposium on Circuits andSystems 212ICECS: IEEE International Conference onElectronics, Circuits and Systems：213ISCAS: International Symposium Circuit andSystem 214RFIC: IEEE Symposium on Radio Frequency Integrated Circuits215ACM RECOMB: Int. Conference on Research inComputational Molecular Biology216IEEE CSB: Computer Society Bioinformatics217International Conference of the IEEE Engineeringin Medicine and Biology Society （EMBC）218PSB: Pacific Symposium on Biocomputing219WABI:Workshop on Algorithms in Bioinformatics220CSB: IEEE Computational Systems BioinformaticsConference221ISMB：Annual International Conference on Intelligent Systems for Molecular Biology222ECCB：European Conference on Computational Biology223APBC：Asia Pacific Bioinformatics Conference224COCOON：Annual International Computing and Combinatorics Conference225CPM：Annual Symposium on Combinatorial Pattern Matching226EMBC：IEEE International Conference of the Engineering in Medicine and Biology Society227Geospatial Information and Technology Association（GITA） Annual Conference228International Geoscience and Remote Sensing Symposium （IGRSS）229International Society for Photogrammetry and Remote Sensing （ISPRS）Technical Commission Symposium230International Conference on Geoinformatics 231IEEE SKG (Semantics, Knowledge and Grid)232WWW: The ACM International World Wide Web Conference233International Semantic Web Conference234ACM SIGMOD: ACM SIGMOD Conf on Management of Data235ACM PODS Conference236VLDB: The ACM International Conference on Very Large Data Bases237IEEE ICDE - International Conference on Data Engineering会议介绍ACM的旗舰会议之一，也是网络领域顶级学术会议，内容侧重于有线网络，每年举办一次，录用率约为10%左右。

z
Intense research activities in universities and other research institutions Drastically increasing number of publications in recent years Large amount of funding by the German government
Model predictive control (MPC)
Modern, optimization-based control technique Successful applications in many industrial fields Can handle hard constraints on states and inputs Optimization of some performance criterion Applicable to nonlinear, MIMO systems
A system is strictly dissipative on a set W ⊆ Z with respect to the supply rate s if there exists a storage function λ such that for all (x , u ) ∈ W it holds that λ(f (x , u )) − λ(x ) ≤ s (x , u ) − ρ(x ) with ρ > 0.
k =0 x (k |t + 1) x (t + 1) state x input u t+1 u (k |t + 1) k =N
Basic MPC scheme